Pre-check: open Task Manager → Performance to see RAM and GPU. Then run this in Command Prompt to see GPU name, driver version and VRAM. Also check free disk space.
nvidia-smi
Do you see your GPU name and "Driver Version"?
Download the Windows installer (OllamaSetup.exe, ~1.9 GB) from ollama.com/download and run it. No admin needed. Needs Windows 10 22H2+ and NVIDIA driver 551.61+.
Do you see the Ollama icon in the system tray?
Open a new Command Prompt or PowerShell and check the version.
ollama -v
Does it print "ollama version is …"?
Run the model — the first run downloads ~2.2 GB.
ollama run phi3:mini
Do you see ">>> Send a message"? Ask anything.
Check GPU use from another terminal: PROCESSOR should show "GPU". Type /bye to leave the chat.
ollama ps
Shows "100% GPU" or a GPU/CPU split?
Pre-check: Task Manager → Performance → GPU shows "AMD Radeon …" and dedicated GPU memory (VRAM). Update the driver with AMD Software: Adrenalin (amd.com).
Can you see the GPU name and VRAM?
Install OllamaSetup.exe (~1.9 GB) from ollama.com/download. Radeon RX 7000 cards use ROCm; other Radeon cards use Vulkan, which is on by default.
Ollama icon in the tray?
Open a new terminal and check the version.
ollama -v
Version printed?
Run the model (~2.2 GB download).
ollama run phi3:mini
See ">>> Send a message"?
GPU check: run ollama ps in another terminal. On iGPU + Radeon laptops, see GGML_VK_VISIBLE_DEVICES in the official docs if it picks the wrong GPU.
ollama ps
PROCESSOR shows GPU?
Pre-check: Apple menu → About This Mac. Note the Chip (Apple M-series or Intel), Memory (RAM) and macOS version.
Noted chip, memory and macOS version?
Download Ollama.dmg from ollama.com/download, open it and drag Ollama to Applications. Needs macOS 14+. Allow it to create the command-line link on first launch.
Ollama icon in the menu bar?
Open Terminal and check the version.
ollama -v
Version printed?
Run the model (~2.2 GB download). M-series Macs use the GPU (Metal) automatically; Intel Macs run on CPU only.
ollama run phi3:mini
See ">>> Send a message"?
Pre-check: check RAM, disk and GPU in a terminal.
free -h
df -h ~
lspci | grep -i -E "vga|3d"
See RAM, free disk and GPU name?
Run the official install script. AMD GPUs need ROCm v7 or Vulkan; NVIDIA needs driver 550+.
curl -fsSL https://ollama.com/install.sh | sh
Script ended with an install-complete message?
Check the version.
ollama -v
Version printed?
Run the model (~2.2 GB download).
ollama run phi3:mini
See ">>> Send a message"?
Check GPU use.
ollama ps
PROCESSOR shows GPU?
Pre-check: see your RAM (Task Manager / About This Mac / free -h). Small tools run without a GPU — just slower.
Know how much RAM you have?
Install exactly as in your OS tab (Windows installer, Mac dmg, Linux script). Without a GPU, Ollama runs on the CPU automatically.
Does ollama -v print a version?
Stick to small models — 1B–3B on 8 GB RAM, only 1B on 4 GB. Run the model (~2.2 GB).
ollama run phi3:mini
Getting answers (even if slowly)?
For speed: close the browser and other apps and keep the laptop plugged in.
Sizes are approximate and can change on the official page.
Errors & fixes
'ollama' is not recognized as an internal or external command
Open a NEW terminal after installing (old ones miss the PATH update). Reinstall if it still fails.
Error: could not connect to ollama app, is it running?
The Ollama server is not running. Windows/Mac: open the Ollama app. Linux: sudo systemctl start ollama, or run ollama serve in another terminal.
Error: model requires more system memory (… GiB) than is available (… GiB)
The model is bigger than your free RAM. Use a smaller model (e.g. llama3.2:1b) and close other apps.
Error: pull model manifest: file does not exist
Wrong model name or tag. Copy the exact name from ollama.com/library (e.g. llama3.2:3b).
ollama ps shows "100% CPU" though I have a GPU
Update the GPU driver (NVIDIA 551.61+ / latest AMD Adrenalin). If the model is bigger than VRAM, Ollama runs it on CPU — pick a smaller one.
Download stuck / very slow / "connection reset"
Network or DNS issue. Restart the router, try a mobile hotspot, or set DNS to 1.1.1.1 / 8.8.8.8. Re-running the same command usually resumes the download.
"No space left on device" / "not enough disk space"
Disk is full. Models are large — keep the free space listed above; clear old downloads and the Recycle Bin.
Phi-3-mini-4k-instruct is a model — it learns new skills through the app you run it in. These add-ons work with Ollama, LM Studio, Open WebUI, Jan, AnythingLLM: give the model tools, web search, files and MCP servers.
Where: Ollama, LM Studio, Open WebUI, Jan, AnythingLLM
ToolsMCPSkillsPlugins
🛡 Stay safe
Read first
Official first — from the AI maker or the service’s own GitHub org / domain.
Read SKILL.md and any scripts before installing. Skills can run code with your permissions.
Never paste base64 or “curl … | sh” download-and-run lines from a skill.
⚠️ MCP servers run with your permissions and can read data you give them; only connect servers you trust. LM Studio asks you to confirm each tool call—keep that on.
Stars, license and last update come from GitHub (17 Nov 2026). Commands are copied from official docs and are not tested on a real machine yet. Access tags are our summary — always check the source. Skills for every AI →